Now that you have the podcasting tools down, put them together in a workflow that keeps you working smoothly, podcast after podcast.
This month, we’re closing out the “Attack of the Pod Penguins” series. In the past 11 months, I’ve tried to pass along everything I can about using open source tools to create your podcast. At the beginning of the series, I made the case that tools such as Audacity, normalize, lame and others are not so different from the tools in the carpenter’s belt. I hope it’s become clearer to you how to use the hammer to drive a nail, the screwdriver to set a screw, and the speed square to assure a 90-degree corner.
But for all the science we’ve covered in the series, creating a compelling podcast is really more art. Once you’re comfortable with the software and hardware tools, pulling it all together with your podcast is really more about the vision and learned ear you carry into the task than the manipulation of the tools. You know what sounds good to you. You have an idea of what will interest your listeners and compel them to spread the word to their friends about your podcast. The tools are really just the means to the end — an interesting, engaging and entertaining few minutes shared with listeners of similar tastes. If the art of creating your podcast isn’t developed and applied liberally, you’re likely to find the experience tedious and frustrating. There’s a reason the phrase “podfade” became popular just a few short months after the genesis of podcasting; if there’s no art in the creation of a podcast, it’s just work, and the podcast fades away. Podcast producers who focus only on the tools and assembly of a podcast quickly lose any joy in the task. And work is really time most budding podcast producers can better spend doing other things. In short, the joy of a podcast is really in the exercise of the art.
In this installment, we’ll work the fuzzy edge between the art and science of podcast production. For me, that edge is the process that allows me to simultaneously implement the tools and exercise the art. Over two and a half years of podcast production, my real liberation has been in the creation of a process that’s as seamless and efficient as possible. The process itself blends the science into the art. It allows me to use the tools without thinking too much about using the tools. Process is the science framework upon which the artistic result is built. When the framework is familiar and efficient, my focus can be on the artistic result.
Over the course of any average week, I’m producing seven podcast editions: three “IndieFeed Blues” podcasts and four distinctly different versions of “The Roadhouse.” On a high-volume week, I’ll also produce as many as 16 editions of the “PRS Podcast” for the American Journal of Plastic and Reconstructive Surgeons and, possibly, a new edition of “The ClarkCast.” That’s 24 podcast files over a seven-day period. It would be impossible to produce podcasts of any quality if I were to focus strictly on the tools. A familiar and comfortable process for each of these podcasts elminates much of the tangible labor, freeing up my efforts to focus on the intangible sound, pacing and overall feel of the podcasts.
So, in this final edition of “Attack of the Pod Penguins,” I’ll share with you the process that works for me in each of these podcasts. While it’s not likely that your process for producing a podcast will exactly mirror my own, hopefully, you’ll glean enough information to adapt these processes to your own work style and to your knowledge of the tools.
Whether you’re producing a music or talk podcast, there is some commonality between processes. At a high level, you’ll find yourself taking similar actions regardless of the podcast type. I’ve always found it useful to execute those actions in a similar order.
1. Gather the source material. 2. Sequence the source material. 3. Write and record any required narrative copy. 4. Sweeten all the source material. 5. Edit, tag and post the audio files. 6. Edit and post the rss feed file.
That’s six high-level steps to producing a podcast. Let’s break them out one by one. Gathering the source material In the music podcasts I produce, gathering and sequencing the source material can be the longest and most taxing step in the process. In the case of both “The Roadhouse” and “IndieFeed Blues,” music is obviously the primary source material. In the case of “The Roadhouse,” I’ve secured permissions from many labels individually to use their music in the show. In most instances, those labels provide promotional CDs — the best possible source material. In the case of “IndieFeed Blues,” permissions have been secured in a more roundabout way. For that podcast, I use music licensed for podcasters through the Independent Online Distribution Alliance (IODA) . IODA works with independent labels to secure licensing agreements. The primary requirement for use of that music in a podcast is reporting the number of downloads of the podcast in which it’s used. IODA then makes those numbers available to the labels to track their promotional efforts. Sequence the source material in sourcing and sequencing music for these podcasts, the length of the podcast itself is important.
As “The Roadhouse” is a long-form podcast coming in at around an hour per edition, mood and pacing are important. In general, I try to set a mood for each edition, opening and closing on a high note with cuts that are uptempo and in a major key. In between, I expend a fair amount of effort to create distinct musical highs and lows, working toward a cut near the end that has a clear “closing time” feel.
“IndieFeed,” on the other hand, is a one-shot podcast — a spoken intro, one song and a spoken outro. The musical sequence is irrelevant. In sourcing “IndieFeed” music, I’m looking for maximum impact. That doesn’t necessarily mean that every cut in “IndieFeed” is uptempo. It just means that, like a short story, every note should deliver a desired effect.
With “The Roadhouse” music selected primarily from CDs, I turn to open source tools to rip the chosen cuts from the disc.
$ cdparanoia -vW [track number] [podcast_track_number-ripped_track_name.wav]
Note than when I’m ripping the cut from CD, I’m naming the resulting file with a name that includes a number indicating where in the show the track will play. This is a quick shorthand way for me to remember what I feel to be the best sequence of music for the show. I follow the same naming convention for those cuts sourced from MP3 files as well, renaming them as necessary. With the show track number at the beginning of the file name, the files will display in the correct show order when listing the source directory. And, in that vein, it’s also important to rename any other files used in the show accordingly. Intros, voiceovers, bumpers and other transitional pieces, ads — all are named or renamed in show sequence order.
As I’ve noted in an earlier installment, I also prefer to work with .wav source files. Some source .mp3 files will require conversion. For that, I turn to a quick little bash script I’ve named convert_mp3.sh:
#!/bin/sh
for i in *.mp3; do
mpg123 -w `basename $i .mp3`.wav $i;
done
This script calls the mpg123 app to write any .mp3 files in a given directory to .wav files.
(Note: There’s no improvement in sound quality when converting from .mp3 to .wav — the quality of the .wav file will never be better than the quality of the original MP3 file. But, in some cases, .wav files are easier to manipulate with other open source tools than .mp3 files. And, you’ll be sure that you’re not reducing the quality of the output file. In the above script, the mpg321 tool will work equally well with the same options.
For housekeeping purposes, I always remove the original .mp3 files after conversion:
$ rm -rf *.mp3
Write and record any required narrative copy This is the point at which the creative side really starts to take over. It’s the way I tie together all the elements of the podcasts. And, in a sense, it’s everything in a spoken-word podcast. In the case of the music podcasts I produce, this copy takes the form of bulleted notes – little snippets of information that provide background on the artists and the labels. For me, it’s important to maintain a relaxed feel for “The Roadhouse” and “IndieFeed,” thus the use of quick notes rather than copy that will be read and recorded word for word. The technique works equally well for spoken-word podcasts, focusing the narrative on the key points to be conveyed while maintaining an informal feel.
For both “The Roadhouse” and “IndieFeed,” the voiceovers are recorded separately from the musical source material. In other words, these are not podcasts that are recorded in a single sitting, start to finish. That’s only because I place a very high value on the transitions between pieces, musical and spoken — they should be as tight as possible. To me, that’s a critical element of the pacing and feel of the podcasts. I’ve never been able to achieve the tight pacing I expect in a live recording setting. With the available tools, it’s more likely that I’ll achieve that pacing by editing the individual elements into a whole. In fact, I’ll often leave verbal stumbles in the final edit. A seamless spoken narrative is less important to me in the music podcasts than the tight fit of all the distinct elements as a whole. Honestly, I’m a freak about those transitions between voice and music, from one song to the next, and from music to imaging pieces and back. I’ve been known to adjust elements of a music podcast by fractions of a second, just to achieve the sound that’s right to my ear.
All of the source material for both “The ClarkCast” and the PRS podcast is recorded elsewhere and provided to me electronically. In the case of “The ClarkCast,” writing the narrative is often more a matter of sequencing critical elements of the raw files into a coherent end-to-end story. I keep copius notes on those elements, using GEdit. After two or three playbacks of the source material, I’ve garnered a feel for the proper sequence of elements and have scripted the segments into what feels like the proper order.
In all cases, consistent with the process of sourcing the material, the files — recorded or edited into distinct elements — are saved and numbered according to their sequence in the final podcast. Sweeten all the source material. It’s very unlikely that your pile of source material will sound exactly as you’d like in the final edit. In all cases, the volume levels will vary from one piece to the next. In the case of voice elements, you may require a bit more presence and resonance. If the voice files are mono recordings, you’ll certainly want to enhance them for stereo. Again, our open source toolkit has everything you’ll need for audio sweetening.
With all the sequence-named files in a single directory, my first sweetening step is to use the normalize tool to bring all elements to a comparable volume level:
$ normalize -m *
This starts a two-step process that a) calculates the average gain from all files and, b) adjusts the gain on each individual file to the average. In short, it gets you close to having a consistent volume level throughout your final podcast.
With the files normalized, I use Audacity for the first time in the editing process. Opening each voice piece in Audacity, I split mono files into stereo by selecting the entire file (Edit -> Select All), copying the original mono (Edit -> Copy), creating a new mono track (Tracks – > Create New), and pasting (Edit – > Paste). The result is two discrete mono tracks that can be manipulated to enhance the overall stereo image of the file.
Before broadening the voice imaging, I bring up the low end in the file (Edit -> Select All, Effects -> Bass Boost). In general, bringing up the 200 Hz frequency by 6 Db using the sliders in the Audacity window will round out the sound of the recorded voice, providing a nice boost in the lower ranges without overpowering the other frequencies.
Finally, I broaden the mono image of the voice file by pushing the pan in the upper track hard left, and the lower track hard right. The voice will remain centered in the stereo image, but will be broader, extending fully into both channels. Save the files, and all the initial sweetening is complete. Edit, tag and post the audio files.
At this point, the bulk of the work is complete. All that’s left is assembling the numbered files into their proper order. This is accomplished by opening a new Audacity window, importing each individual file and sliding it around in the window to create your transitions. In the case of “The Roadhouse,” I always assemble the show in segments, each ending with the spoken narrative. This allows me to mix and match musical groups if, in the end, they make more sense reordered. It also allows me to overcome a slight drawback in Audacity; stacking more than four or five tracks in a single window can cause problems when trying to slide the tracks around. Occasionally, a track will simply disappear when moved. This can be corrected by selecting Edit -> Undo Adjust Timeline, but it’s easier to avoid the problem altogether by limiting the number of tracks stacked in a window.
With the segments edited into their chosen order, export the full file to a .wav by selecting Edit -> Export As -> .wav. Depending on file size and processor speed, this may take as long as 5 to 6 minutes to complete.
Consistent with the use of open source tools to complete individual tasks, my final .mp3 is encoded with lame, even though Audacity offers an Export to .mp3 option. The command line lame tool offers complete flexibility in creating an .mp3 file, including bitrate, comments and some rudimentary tagging. The lame command for a typical Roadhouse looks like this:
$ lame -b 160 -q 0 –noreplaygain –ta “Tony Steidler-Dennison” –tt “Roadhouse Premium 118” –tl “The Roadhouse Premium” –ty 2007 –tc “Roadhouse Premium 118 – Freewheelin’ Feel” roadhouse_premium_118.wav roadhouse_premium_118.mp3
This command encodes the file at 160 Kbps (-b 160) leaves the file gain intact (–noreplaygain), writes the basic ID3 tags (–ta [artist], –tt [title], –tl [album], –ty [year], –tc [comments]), and writes the input file (roadhouse_premium_118.wav) to the output file (roadhouse_premium_118.mp3).
Note that the genre “Podcast” isn’t written in the lame command. Lame follows the ID3 standard regarding genres. “Podcast” is not yet one of those recognized genre names. To tag the genre, I use EasyTag. With that task accomplished, the file is uploaded to the server.
Edit and post the rss feed file The real beauty of podcasting is the marriage of rss to the media file. Several elements of the rss file you’ll create for your podcast are required, while several others are optional — including those that will make your podcast available in iTunes. Again, by way of example, here’s a typical example of a single rss entry for “The Roadhouse.”
You can really see the beauty of XML in this entry. Every element of the podcast metadata is easy to understand. Posted to a server, you’re ready to count your new subscriber and download numbers.
That covers what I’ve found to be a sensible workflow for creating nearly any podcast. It’s one that has established enough of a routine to separate the technical work from the creative. As I’ve said, that fuzzy edge is where the technical meets the artistic.
And, that closes out the “Attack of the Pod Penguins” series. Thanks for sticking around for nearly a year. I hope the tips, theory and practical advice will help you accomplish the goal of nearly any open source podcaster: creating compelling audio content, finding an audience for that content, and contributing to the overall use of open source tools.
